huggingface/candle
1
0
Fork 0
mirror of https://github.com/huggingface/candle.git synced 2025-06-19 03:54:56 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add a matmul cuda example.

Browse Source
This commit is contained in:
laurent
2023-06-22 19:44:26 +01:00
parent 0671b8c369
commit aebffcfc13
2 changed files with 11 additions and 7 deletions
Download Patch File Download Diff File Expand all files Collapse all files

9
src/cuda_backend.rs
View File

@ -159,14 +159,15 @@ fn gemm_config<T>(
) -> StridedBatchedConfig<T> {
// https://docs.nvidia.com/cuda/cublas/index.html#cublas-t-gemm
use cudarc::cublas::sys::cublasOperation_t;
println!("{:?} {:?} {:?}", lhs_stride, rhs_stride, (b, m, n, k));
let gemm = GemmConfig {
alpha,
beta,
m: m as i32,
n: n as i32,
k: k as i32,
lda: lhs_stride[lhs_stride.len() - 2] as i32,
ldb: rhs_stride[rhs_stride.len() - 2] as i32,
lda: m as i32,
ldb: k as i32,
ldc: m as i32,
transa: cublasOperation_t::CUBLAS_OP_N,
transb: cublasOperation_t::CUBLAS_OP_N,
@ -174,8 +175,8 @@ fn gemm_config<T>(
StridedBatchedConfig {
batch_size: b as i32,
gemm,
stride_a: lhs_stride[0] as i64,
stride_b: rhs_stride[0] as i64,
stride_a: (m * k) as i64,
stride_b: (n * k) as i64,
stride_c: (m * n * k) as i64,
}
}